PROJECT SUMMARY Cancer is a leading cause of death in the United States, with a third of diagnoses attributed to modifiable risk factors including obesity and poor diet. Understanding the role of these risk factors in cancer development is crucial for the provision of appropriate public health guidance for cancer prevention. Breast cancer remains the highest incidence cancer among women in the United States, and incidence rates for endometrial cancer are projected to dramatically increase over the next decade. Epidemiological, clinical and laboratory evidence suggests that diet may be relevant for breast and endometrial cancer prevention, although the evidence base lacks consistency, perhaps owing to imprecise dietary measures, and a failure to account for associations that may vary by tumor subtype. Furthermore, endometrial cancer is especially obesity-driven, but the underlying mechanisms have yet to be fully characterized. There is a need for better objective measures of diet, including dietary biomarkers, that can be used to improve dietary assessment. Metabolomics is a novel and emerging technology in molecular epidemiology that can be used to measure hundreds to thousands of circulating metabolites simultaneously, more than 200 of which have been recently linked to an individual?s diet and/or adiposity. Furthermore, metabolomics can be used to highlight biological mechanisms of interest in relation to disease. This novel technology is advancing rapidly, and much work is yet to be done in applying it in an epidemiologic context. To address this unmet need, I propose to apply metabolomics to understanding the relationships of diet and adiposity with female cancers. I will first quantify the relationship between circulating metabolites and habitually consumed foods in a feeding study with a gold-standard measure of diet (weighed food) among postmenopausal women in order to develop objective dietary biomarkers. Such work is critical for correct interpretation of any diet-related metabolite signals that may be observed in subsequent cancer studies. Next, I will measure the association between pre-diagnostic circulating diet-related metabolites and estrogen receptor- negative (ER-) breast cancer using nested case-control data from three prospective cohorts. ER- breast cancer is rare, aggressive and understudied. Consequently, the etiology is poorly understood, including potential dietary risk factors. Finally, I will determine whether pre-diagnostic circulating metabolites are associated with incident endometrial cancer, and their relation to diet or adiposity using nested case-control data from four prospective cohorts. No studies have explored metabolite signatures of endometrial cancer and their relation to adiposity and diet using data from US-based prospective studies. These studies may uncover unknown metabolic pathways involved in the etiology of breast and endometrial cancer that may also be applicable to other obesity-driven cancers, and identify key pathways for developing and evaluating targeted cancer prevention interventions.